The study was conducted to cultivate the understanding of different weather perceptions as well as indicators by rice farmers of Mbarali and Kyela districts in Mbeya region of Tanzania. Primary and secondary data were obtained from the research area and Tanzania Meteorological Agency. Garret technique and Multipredictor logistic regression were applied for data analyses. The result shows that farmers have different perceptions on various weather variables like rainfall, temperature, drought, wind, and floods. Likewise with logistic regression is affirmed that awareness towards weather variations increased by 0.60 times and 2.0 times if farmer’s education and extended floods each increase one times respectively. In addition, various weather variability indicators like lack of enough rainfall, increase in temperature, drying of tender tree leaves, and severe sunshine were identified. More so, factors hampering immediate local adaptations to weather variability were also identified. Authors recommended on establishment of a working coordination mechanism by incorporating local perceptions, indicators, and adaptions to rescue rural communities from the ongoing food insecurity threats.